IMPORTANCE Previous studies that have shown tai chi to improve sleep were mainly based on subjective assessments, which might have produced results confounded by self-reporting bias. OBJECTIVE To compare the effectiveness of tai chi for improving sleep in older adults with insomnia with conventional exercise and a passive control group using actigraphy-based objective measurements. DESIGN, SETTING, AND PARTICIPANTS This randomized, 3-arm, parallel group, assessor-masked clinical trial was conducted at a single research unit in Hong Kong between August 2014 and August 2018. Eligible participants, aged 60 years or older and with chronic insomnia, were randomly allocated into tai chi training, exercise, and control groups. INTERVENTIONS 12-week tai chi training, 12-week conventional exercise, and no intervention control. MAIN OUTCOMES AND MEASURES Primary outcomes were measures taken from actigraphy sleep assessment. Secondary outcomes included remission of insomnia, insomnia treatment response, Pittsburgh Sleep Quality Index score, Insomnia Severity Index score, and self-reported sleep using a 7-day sleep diary. Assessments were performed at baseline, end of the intervention (postintervention), and 24 months after the intervention (follow-up). Data analysis was performed from September 2018 to August 2020. RESULTS A total of 320 participants (mean [SD] age, 67.3 [6.8] years; mean [SD] insomnia duration, 124.4 [134.5] months; 256 [80.0%] women) were randomly allocated into control (110 participants), exercise (105 participants), and tai chi (105 participants) groups and included in the data analysis. Compared with the control group, the exercise and tai chi groups showed improved sleep efficiency (exercise vs control: adjusted mean difference, +3.5%; 95% CI, 1.8-5.2; P < .001; tai chi vs control: adjusted mean difference, +3.4%; 95% CI, 1.6-5.1; P < .001) and reductions of wake time after sleep onset (exercise vs control: -17.0 minutes; 95% CI, -24.9 to -9.0; P < .001; tai chi vs control: -13.3 minutes; 95% CI, -21.3 to -5.2; P = .001) and number of awakenings (exercise vs control: -2.8 times; 95% CI, -4.0 to -1.6; P < .001; tai chi vs control: -2.2 times; 95% CI, -3.5 to -1.0; P < .001) as assessed by actigraphy at postintervention; although there were no significant differences between the exercise and tai chi groups. The actigraphy-assessed beneficial effects were maintained in both intervention groups at follow-up. CONCLUSIONS AND RELEVANCE Conventional exercise and tai chi improved sleep and the beneficial effects sustained for 24 months, although the absolute improvements in sleep parameters were modest. Improvements in objective sleep parameters were not different between the tai chi and exercise groups, suggesting that tai chi can be an alternative approach for managing insomnia.